A few days ago, I reported a problem I started seeing immediately after
applying RH5.0 updates as of April 3 (probably a coincidence!?!?, but
shows up in the logs right after upgrades + reboot). Problem was:
kernel: attempt to access beyond end of device
kernel: 03:08: rw=0, want=1634493797, limit=1028128
I have now found the offending file: /var/log/lastlog
Sure enough, if I cp that file, "cp: lastlog: Input/output error" and
same errors reported to /var/log/messages. Now, when I force an fsck on
this partition (e2fsck -vfn /dev/hda8), I get a slew of these
Illegal block # messages. The other thing you'll see is quite a few
files in /var/log that share duplicate block numbers.
Illegal block #78 (544370753) in inode 242807. IGNORED.
Illegal block #79 (824193824) in inode 242807. IGNORED.
Illegal block #80 (808729137) in inode 242807. IGNORED.
....
Illegal block #267 (842282800) in inode 242807. IGNORED.
Inode 242807, i_blocks is 8, should be 6. Fix? no
Duplicate blocks found... invoking duplicate block passes.
Pass 1B: Rescan for duplicate/bad blocks
Duplicate/bad block(s) in inode 242806: 993214 993215 993216 993217
Duplicate/bad block(s) in inode 242807: 988480
Illegal block number passed to ext2fs_test_block_bitmap #824193824 for multiply
claimed block map
Illegal block number passed to ext2fs_test_block_bitmap #808729137 for multiply
claimed block map
....
Duplicate/bad block(s) in inode 242824: 993202 993203 993204 993205 993206 993207
993208
Duplicate/bad block(s) in inode 242826: 993202 993203 993204 993205 993206 993207
993208 993214 993214 993215 993216 993217 993220 993220
Duplicate/bad block(s) in inode 242828: 988480
Pass 1C: Scan directories for inodes with dup blocks.
File /log/maillog (inode #242828, mod time Tue Apr 21 14:00:18 1998)
has 1 duplicate block(s), shared with 1 file(s):
/log/lastlog (inode #242807, mod time Tue Apr 21 12:13:53 1998)
File /log/messages (inode #242826, mod time Tue Apr 21 14:00:05 1998)
has 14 duplicate block(s), shared with 2 file(s):
/log/cron (inode #242824, mod time Tue Apr 21 14:00:00 1998)
/log/wtmp (inode #242806, mod time Tue Apr 21 12:46:58 1998)
File /log/cron (inode #242824, mod time Tue Apr 21 14:00:00 1998)
has 7 duplicate block(s), shared with 1 file(s):
/log/messages (inode #242826, mod time Tue Apr 21 14:00:05 1998)
File /log/lastlog (inode #242807, mod time Tue Apr 21 12:13:53 1998)
has 1 duplicate block(s), shared with 1 file(s):
/log/maillog (inode #242828, mod time Tue Apr 21 14:00:18 1998)
File /log/wtmp (inode #242806, mod time Tue Apr 21 12:46:58 1998)
has 4 duplicate block(s), shared with 1 file(s):
/log/messages (inode #242826, mod time Tue Apr 21 14:00:05 1998)
Now my question. Is there anything I should do to track down the cause of
the failure before letting e2fsck fix this mess. I've _NEVER_ had
anything like this happen on my Linux machines, so this strikes me as
being peculiar. I don't believe it is an indication of a bad sector on
the hard drive (I would expect that the kernel would report data read
errors instead). Since these problems are proliferating in the log files,
could "logrotate" have started doing something to me over the last few
weeks?
Last, in running the same e2fsck on my / partition, "Pass 5: Checking group
summary information" reports many, many:
Block bitmap differences: -92319 -92320 -92321 -92322 -92323 -92324 ...
What is this and is it indicating another problem?
Any suggestions would be appreciated.
Thanks,
Mike
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
